DBS-C01 Management and Operations — 20 Questions

Question 1

A company runs a production Amazon RDS for PostgreSQL DB instance. The database experiences intermittent high latency during peak hours. The company's monitoring shows that CPU utilization is below 40%, memory is adequate, and network throughput is normal. The DB instance uses General Purpose SSD (gp2) storage with 3,000 Provisioned IOPS. Which TWO actions should a database specialist take to diagnose and resolve the performance issue? (Choose two.)

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Check the Burst Balance metric in Amazon CloudWatch for the DB instance.. Gp2 storage uses a credit-based burst model, and the Burst Balance metric indicates the remaining I/O credits. If the burst balance is depleted, the instance is throttled to the baseline IOPS (3,000 for this size), causing high latency despite low CPU and adequate memory. Option E is also correct because monitoring Read/Write Latency metrics in CloudWatch directly measures storage performance and can confirm whether high latency is due to storage I/O issues. Together, these two metrics help identify if the gp2 burst credits are exhausted or if other storage-related latencies are occurring. Options B, C, and D are not the most appropriate first diagnostic steps: B suggests changing storage type without diagnosis, C does not directly address storage I/O, and D checks memory pressure, which is not indicated given adequate memory.

Answer

Change the storage type to Provisioned IOPS SSD (io1) with 10,000 IOPS.

Answer

Increase the Read IOPS by modifying the instance to a larger instance class.

Answer

Review the SwapUsage metric for the DB instance to check for memory pressure.

Question 2

Which TWO actions can help reduce the recovery time objective (RTO) for an Amazon RDS for PostgreSQL DB instance in the event of a failure? (Choose 2.)

Accepted Answer

Enable Multi-AZ deployment.. Enabling Multi-AZ deployment for Amazon RDS for PostgreSQL provides automatic failover to a standby replica in a different Availability Zone. This reduces RTO by automating the DNS change and database promotion, typically completing within 60–120 seconds without manual intervention. Option C is correct because creating a read replica in a different Availability Zone allows you to manually promote it to a standalone instance during a failure, which can be faster than restoring from a backup, thereby reducing RTO. Although promotion is manual and takes a few minutes, it still offers a lower RTO compared to backup-based recovery. Option D, configuring automated backups with point-in-time recovery, primarily improves recovery point objective (RPO) rather than RTO, as restoring from a backup is typically slower and involves more steps. Options A and E do not directly reduce RTO since increasing instance class size does not aid failover, and manual snapshots also rely on restoration.

Answer

Increase the DB instance class size.

Answer

Configure automated backups with point-in-time recovery.

Answer

Take manual snapshots every hour.

Question 3

A company's Amazon RDS for MySQL instance is experiencing high CPU utilization. The company's application is write-heavy. The database administrator wants to reduce the load on the primary instance. Which solution is MOST effective?

Accepted Answer

Increase the instance size of the RDS instance.. Increasing the instance size (Option A) provides more CPU and IO capacity to handle the write-heavy workload, thereby reducing CPU utilization. Read replicas (B) can offload read traffic but not write load. Multi-AZ (C) enhances availability but does not distribute writes; writes continue on the primary. ElastiCache (D) caches reads, not writes, so it does not reduce write-related CPU load.

Answer

Create a read replica and redirect read traffic to it.

Answer

Enable Multi-AZ deployment to distribute writes.

Answer

Use Amazon ElastiCache to cache frequent writes.

Question 4

A company wants to monitor the performance of an Amazon RDS for MySQL instance. Which TWO metrics are available in Amazon CloudWatch by default? (Choose 2.)

Accepted Answer

FreeableMemory. CPUUtilization (Option B) and FreeableMemory (Option A) are both default CloudWatch metrics for Amazon RDS. CPUUtilization measures the percentage of CPU capacity used by the database instance. FreeableMemory reports the amount of available random access memory. Both are automatically published to CloudWatch every minute without additional setup. DatabaseConnections (Option E) is also a default metric, but in the context of this question, the two correct answers are CPUUtilization and FreeableMemory. QueryThroughput (Option C) is not a standard CloudWatch metric for RDS; it is available from Performance Insights. ReadThrottleEvents (Option D) is not a default metric; it is specific to DynamoDB or RDS with provisioned IOPS. Therefore, the correct answers are A and B.

Answer

QueryThroughput

Answer

ReadThrottleEvents

Answer

DatabaseConnections

Question 5

A database administrator runs the above CLI command. The DB instance has pending modifications. Which modification will take effect immediately without requiring a reboot?

Accepted Answer

BackupRetentionPeriod. The `--apply-immediately` flag causes eligible modifications to be applied immediately. Both `BackupRetentionPeriod` and `MasterUserPassword` are parameters that can be changed dynamically without requiring a reboot. `BackupRetentionPeriod` adjusts the backup schedule in the background, and `MasterUserPassword` updates the password instantly via the RDS API. In contrast, `DBInstanceClass` and `AllocatedStorage` require a reboot even with `--apply-immediately`.

Answer

DBInstanceClass

Answer

AllocatedStorage

Question 6

A database specialist sees the above output for an RDS MySQL instance. The application reports that the instance is not accepting writes. What is the most likely cause?

Accepted Answer

The instance is a read replica and cannot accept write requests.. The most likely cause is that the RDS MySQL instance is configured as a read replica. Read replicas in Amazon RDS are designed to handle read traffic only and do not accept write operations. Any attempt to write to a read replica will be rejected by the database engine, which matches the symptom described.

Answer

The allocated IOPS are insufficient for the write workload.

Answer

Multi-AZ is not enabled, so the instance cannot accept writes.

Answer

Storage encryption is enabled, which prevents write operations.

Question 7

A company is using a Multi-AZ RDS for MySQL instance. The primary instance recently experienced an unexpected failover, and the application experienced a brief interruption. The database administrator wants to investigate the root cause. Which AWS service should be used to review the failover event and its details?

Accepted Answer

Amazon RDS Event Notifications. Amazon RDS Event Notifications publish events related to DB instances, including failovers. They provide details such as the event source, message, and time, enabling root cause analysis. While AWS CloudTrail can record API calls that trigger failovers (e.g., RebootDBInstance with Failover), automatic failovers due to underlying issues are not captured as API calls. AWS Config tracks configuration changes, and AWS Trusted Advisor offers best practice checks but does not log specific events. Therefore, to review the failover event and its details, RDS Event Notifications is the most appropriate service.

Answer

AWS Config

Answer

AWS Support (Trusted Advisor)

Answer

AWS CloudTrail

Question 8

A company is running Amazon RDS for Oracle with a Multi-AZ deployment. The database is experiencing performance degradation due to high I/O waits. The team suspects that the issue is related to the transaction log (redo log) writes. Which THREE metrics should be monitored in Amazon CloudWatch to diagnose the issue?

Accepted Answer

DiskQueueDepth. Options A, C, and E are correct because high redo log sync time, high redo log file sync waits, and high disk queue depth are indicators of I/O bottlenecks affecting redo log writes. Option B is wrong because CPU utilization does not directly indicate redo log issues. Option D is wrong because read latency is not directly related to redo log writes.

Answer

CPUUtilization

Answer

ReadLatency

Question 9

A company is designing a disaster recovery plan for an Amazon RDS for Oracle DB instance. The primary region is us-east-1, and the DR region is us-west-2. The RPO must be less than 5 minutes, and the RTO must be less than 15 minutes. Which THREE actions should be taken? (Choose THREE.)

Accepted Answer

Create a cross-region read replica in us-west-2.. Amazon RDS for Oracle supports cross-region read replicas, which use asynchronous replication to keep the replica in us-west-2 within seconds of the primary. This can achieve an RPO of less than 5 minutes and, with manual promotion, an RTO of less than 15 minutes. Option B is correct because enabling cross-region automated backups copies backups to us-west-2, providing a baseline backup with a longer retention period. While automated backups alone cannot achieve a 5-minute RPO, they are a critical part of a comprehensive DR strategy, offering protection against data corruption and enabling point-in-time recovery. Option D is correct because AWS DMS with ongoing replication (change data capture) can replicate transactions from the source Oracle DB to a target in us-west-2 with sub-second latency, achieving the required RPO and RTO when combined with a pre-provisioned target.

Answer

Deploy a Multi-AZ standby in us-east-1.

Answer

Take daily manual snapshots and copy them to us-west-2.

Question 10

A company uses Amazon DynamoDB with global tables in a multi-Region active-active configuration. The application writes to the table in us-east-1 and reads from us-west-2. Recently, a network partition occurred between the two regions, and after recovery, some data was lost. What is the most likely reason for the data loss?

Accepted Answer

The last-writer-wins conflict resolution used timestamp-based coordination without clock synchronization.. Global tables use last-writer-wins (LWW) conflict resolution based on timestamps. If clocks are not synchronized, an older write from one region can overwrite a newer write. Option A is incorrect because eventually consistent reads do not cause data loss; they may return stale data but not overwrite. Option B (LWW with unsynchronized clocks) is the correct cause. Option C is incorrect because TTL expiration would affect data in both regions, not cause loss after partition. Option D is incorrect because replication resumes after partition recovery, and replication does not stop permanently.

Answer

The application used eventually consistent reads, causing stale data to be overwritten.

Answer

The table had a TTL attribute that expired the data before replication completed.

Answer

The network partition caused the replication to stop, and data was not replicated.

Question 11

A database administrator sees repeated 'Aborted connection' warnings in the RDS for MySQL error log. The application team reports intermittent connection errors. What is the most likely cause of these aborted connections?

Accepted Answer

The application is closing the database connection abruptly without completing the handshake.. The 'Aborted connection' warning in RDS for MySQL typically occurs when a client fails to complete the initial handshake or disconnects abruptly before the authentication completes. This is often caused by the application closing the connection prematurely or a network interruption during the connection setup. Option A is incorrect because wait_timeout affects idle connections, not aborted connections during handshake. Option B is incorrect because a permission error would result in 'Access denied' rather than 'Aborted connection'. Option C is incorrect because memory pressure would cause performance degradation, not specifically these warnings.

Answer

The wait_timeout parameter is too low, causing idle connections to be terminated.

Answer

The user 'appuser' does not have permission to connect from host '10.0.0.50'.

Answer

The database instance is running out of memory, causing connection drops.

Question 12

A database specialist is configuring an Amazon RDS for PostgreSQL instance with automated backups enabled. The backup retention period is set to 7 days. The instance experiences a failure and becomes unavailable. What is the maximum amount of time the specialist has to recover the database to the most recent possible point in time?

Accepted Answer

7 days. With automated backups enabled and a retention period of 7 days, Amazon RDS for PostgreSQL allows point-in-time recovery to any point within the backup retention period, up to the latest restorable time (typically within the last 5 minutes). Therefore, the maximum time window for recovering to the most recent possible point in time is 7 days. Option A is incorrect because 1 hour is too short; the retention period covers up to 7 days. Option C is incorrect because 14 days exceeds the configured retention period of 7 days. Option D is incorrect because 35 minutes is not a standard limitation for point-in-time recovery; the restriction is the retention period.

Answer

1 hour

Answer

14 days

Answer

35 minutes

Question 13

A company is running an Amazon RDS for MySQL Multi-AZ DB instance. The primary instance in us-east-1a experiences an unexpected failure. After the automatic failover, the application team reports that write latency has increased significantly. The new primary instance is in us-east-1b. The DB instance class and storage configuration are identical. What is the MOST likely cause of the increased write latency?

Accepted Answer

The application is connecting to the DB instance in a different Availability Zone, increasing network latency.. After failover, the new primary DB instance resides in us-east-1b, while the application likely continues to connect to the original endpoint or is still running in us-east-1a. This cross-AZ network hop introduces additional latency for write operations, as the application must send data over the network between Availability Zones. The DB instance class and storage are identical, so performance differences are not due to hardware changes.

Answer

The DB instance class in us-east-1b is a different size than the original.

Answer

The Multi-AZ configuration uses asynchronous replication, causing higher latency.

Answer

The new primary is in the same Availability Zone as the application, but the standby is in a different AZ, causing synchronous replication overhead.

Question 14

A database administrator is troubleshooting an Amazon RDS for PostgreSQL DB instance that is experiencing high CPU utilization. The administrator runs the following query to find the current running queries:

SELECT pid, now() - pg_stat_activity.query_start AS duration, query, state
FROM pg_stat_activity
WHERE state = 'active';

The output shows a high number of queries with a state of 'active' and durations exceeding several minutes. What should the administrator do FIRST to reduce CPU utilization?

Accepted Answer

Use pg_terminate_backend to terminate the long-running queries.. The immediate cause of high CPU utilization is the long-running active queries consuming resources. Using pg_terminate_backend to terminate these queries will quickly free up CPU cycles, providing immediate relief. This is the first troubleshooting step before making configuration changes or scaling, as it directly addresses the symptom shown in the pg_stat_activity output.

Answer

Modify the max_connections parameter to limit concurrent sessions.

Answer

Scale up the DB instance to a larger instance class.

Answer

Enable pg_stat_statements to collect query performance data.

Question 15

A company is using Amazon DynamoDB with on-demand capacity mode. The application experiences occasional throttling during peak hours. The operations team wants to reduce throttling without changing the application code. What should they do?

Accepted Answer

Switch to provisioned capacity mode and configure auto scaling.. DynamoDB on-demand capacity mode does not allow manual adjustment of capacity units; it scales automatically but can still throttle if traffic exceeds the previous peak by a large margin. Switching to provisioned capacity mode with auto scaling allows you to set a higher minimum capacity and scale proactively based on actual usage patterns, reducing throttling without code changes. This approach gives more control over capacity limits while still automating adjustments.

Answer

Increase the read and write capacity units in the on-demand mode.

Answer

Enable DynamoDB Accelerator (DAX) to cache reads.

Answer

Use DynamoDB global tables to distribute traffic across regions.

Question 16

A company is migrating an on-premises Microsoft SQL Server database to Amazon RDS for SQL Server. The database uses SQL Server Agent jobs, custom CLR assemblies, and cross-database queries. Which of the following will require modification before migration?

Accepted Answer

SQL Server Agent jobs. SQL Server Agent jobs are not supported in Amazon RDS for SQL Server because RDS is a managed service that does not provide access to the underlying operating system or the SQL Server Agent service. To migrate job scheduling, you must use alternatives such as AWS Database Migration Service (DMS) tasks, AWS Lambda, or Amazon RDS for SQL Server native scheduling via stored procedures and Windows Task Scheduler on an EC2 instance.

Answer

Cross-database queries

Answer

Custom CLR assemblies

Answer

Stored procedures that use dynamic SQL

Question 17

A company is using Amazon RDS for MySQL with a cross-Region read replica to support disaster recovery. The primary DB instance is in us-west-2, and the read replica is in us-east-1. The read replica is used for reporting and also serves as a failover target. The operations team notices that the read replica lag is consistently above 10 seconds during peak hours. What should the team do to reduce replica lag?

Accepted Answer

Increase the DB instance class of the read replica.. Increasing the DB instance class of the read replica provides more CPU and memory resources, which allows the replica to apply changes from the binary log more quickly. Cross-Region replication lag is often caused by the replica being unable to keep up with the write rate on the primary, so scaling up the replica directly addresses the bottleneck in applying binlog events.

Answer

Enable Multi-AZ on the primary DB instance.

Answer

Increase the backup retention period for the primary DB instance.

Answer

Disable binary logging (binlog) on the primary DB instance.

Question 18

A company is using an Amazon Aurora MySQL DB cluster. The company wants to implement a backup strategy that supports point-in-time recovery (PITR) with a recovery time objective (RTO) of 15 minutes and a recovery point objective (RPO) of 5 minutes. Which TWO actions should the company take?

Accepted Answer

Configure automated backups with a retention period of at least 1 day.. Aurora automated backups with a retention period of at least 1 day provide continuous backup and point-in-time recovery (PITR) with an RPO of 5 minutes because transaction logs are uploaded every 5 minutes. This meets the required RPO of 5 minutes, and restoring can achieve an RTO under 15 minutes. Option E is correct because Aurora Backtrack allows rewinding the cluster to a specific point in time, achieving an RPO of 5 minutes and an RTO of minutes (typically under 15). No other option meets both the stated RTO and RPO.

Answer

Create manual snapshots every 5 minutes.

Answer

Enable cross-Region replication for the cluster.

Answer

Enable parallel query for the cluster.

Question 19

A company is running a production Amazon DynamoDB table with on-demand capacity. The table experiences occasional throttling during traffic spikes. The table's partition key is a timestamp, and the workload is write-heavy. The operations team needs to reduce throttling. Which THREE actions should the team take? (Choose three.)

Accepted Answer

Add a random suffix to the partition key to distribute writes evenly.. Adding a random suffix to the timestamp partition key breaks the sequential write pattern, distributing writes evenly across all partitions. This prevents hot partitions, which are the root cause of throttling in a write-heavy workload with a monotonically increasing partition key.

Answer

Increase the read capacity units for the table.

Answer

Enable DynamoDB global tables to distribute write traffic.

Question 20

A database administrator runs the above AWS CLI command to troubleshoot replication issues. The DB instance 'mydb' is a read replica of 'my-source-db'. The administrator notices that the replica lag is increasing. Which of the following is the MOST likely cause?

Accepted Answer

The read replica is in a different AWS Region than the source.. When a read replica is in a different AWS Region than the source, the replication traffic must traverse the public internet or a VPN connection, introducing network latency and potential bandwidth constraints. This cross-region lag is a common cause of increasing replica lag, as the asynchronous MySQL replication relies on a single I/O thread to download the binary log events from the source, and any network delay directly impacts the replica's ability to keep up.

Answer

The source DB instance is running a different MySQL version.

Answer

The read replica has Multi-AZ disabled.

Answer

The read replica is using a smaller instance class than the source.